An Approach to Optimize Local Trust Algorithm in SureMsg Service

SureMsg is a trustworthy email service that is based on the instant messaging (IM) platform and which scores the reputation level of email sender by both: global reputation value and local trust value. As compared to the global reputation, the local trust value is more reasonable and precisely represents a sender\´s behaviour in the social network. This paper proposes an approach that is underpinned by the "six- degree-of-separation" to optimize the local trust algorithm in SureMsg so as to decrease the complexity of computation to the local trust value and to obtain a precise score. Experiments are presented that demonstrate that the six-degree-of-separation theory is practicable in SureMsg. Moreover, we show that decreasing the degree from six to two or to three can also obtain a precise score and also decrease the complexity of computation.